dajobe: Wed 10th 7:30pm room 108B
DanC_jam: Presenter(s): Mark R. Crawford, XML Lead, Logistics Management Institute, United States
DanC_jam: argues for UBL as a hub format
DanC_jam: "UBL on track to become international standard for trade through ISO TC 154"
DanC_jam: argues UBL provides the XML payload for ebXML
DanC_jam: non-proprietary, though based on work by CommerceOne
Jhendler: bio of presented (from last year) at presenters bio
DanC_jam: hmm... connect these guys to the Semantic Web Best Practices WG?
DanC_jam: hmm... seem to have a different approach to namespaces
DanC_jam: code lists... controlled vocabularies... definitely a topic for SemWeb best practices
DanC_jam: argues for UBL as a hub format
DanC_jam: "UBL on track to become international standard for trade through ISO TC 154"
DanC_jam: argues UBL provides the XML payload for ebXML
DanC_jam: non-proprietary, though based on work by CommerceOne
Jhendler: bio of presented (from last year) at presenters bio
DanC_jam: hmm... connect these guys to the Semantic Web Best Practices WG?
DanC_jam: hmm... seem to have a different approach to namespaces
DanC_jam: code lists... controlled vocabularies... definitely a topic for SemWeb best practices
danbri: Author's Official Page for O'Reilly XForms Essentials
danbri: Micah's talking on this now, mentioned materials are in his O'Reilly book which is online under an open license.
danbri: See also xforms example page.
danbri: Talk mentions Microsoft's InfoPath; apparently its data schemas recently released (presumably on this MS site somewhere).
danbri: Micah's talking on this now, mentioned materials are in his O'Reilly book which is online under an open license.
danbri: See also xforms example page.
danbri: Talk mentions Microsoft's InfoPath; apparently its data schemas recently released (presumably on this MS site somewhere).
Ankh: Liam's RDF adventure game (might not be 100% working because of bitrot but looks OK at a glance)
dajobe: worst web site recently seen. The only text in this page is in the <title>
dajobe: thanks to Haddock and the award for most ironic use of giftext goes to...
dajobe: lynx -dump -nolist http://www.esws2004.org/|wc -w answer: 1
dajobe: thanks to Haddock and the award for most ironic use of giftext goes to...
dajobe: lynx -dump -nolist http://www.esws2004.org/|wc -w answer: 1
evlist: Two big things in XML now: web services if you want to make money, blog if you're a birk...
DanC_jam: "The Birks: XML is for Blogging"
DanC_jam: DM says there's an incredible amount of data available; U.S. govt now publishes their stuff. Finnish pilots can get more about finland from the U.S. govt than from their own govt
DanC_jam: "... taking a clue from CEOs, who never read anything, but only summaries of things, like we all do for W3C specs." 1/2 ;-)
DanC_jam: the Irish invented putting spaces beteween words.
DanC_jam: The RDF 1.0 spec was the scariest thing ever to come out of W3C. WG didn't use words that anybody understood.
danbri: "RDF stuff good but a bit academic... but some crossover... the bloggers are coming up with stuff like Friend of a Friend, ...."
evlist: DM complains that RDF isn't friendly toward denormalization (can't add denormalized information in reference to resources)
evlist: DM Proposes a "PubXS" format that allows some level of denormalization.
evlist: The "pubxs:link" attribute looks more like xlink:href than like rdf:resource.
logger: See discussion
DanC_jam: "2 special attributes and no mixed content" -- RDF 2.0 in a nutshell
DanC_jam: I wonder if PubXS is isomorphic to XIOR
danbri: his 'ref' thing is basically rdfs:seeAlso, except he defines it as 'better info available from ...' rather than 'maybe some further info available ...'
DanC_jam: Megginson beats the "worse is better" drum pretty hard
DanC_jam: "The Birks: XML is for Blogging"
DanC_jam: DM says there's an incredible amount of data available; U.S. govt now publishes their stuff. Finnish pilots can get more about finland from the U.S. govt than from their own govt
DanC_jam: "... taking a clue from CEOs, who never read anything, but only summaries of things, like we all do for W3C specs." 1/2 ;-)
DanC_jam: the Irish invented putting spaces beteween words.
DanC_jam: The RDF 1.0 spec was the scariest thing ever to come out of W3C. WG didn't use words that anybody understood.
danbri: "RDF stuff good but a bit academic... but some crossover... the bloggers are coming up with stuff like Friend of a Friend, ...."
evlist: DM complains that RDF isn't friendly toward denormalization (can't add denormalized information in reference to resources)
evlist: DM Proposes a "PubXS" format that allows some level of denormalization.
evlist: The "pubxs:link" attribute looks more like xlink:href than like rdf:resource.
logger: See discussion
DanC_jam: "2 special attributes and no mixed content" -- RDF 2.0 in a nutshell
DanC_jam: I wonder if PubXS is isomorphic to XIOR
danbri: his 'ref' thing is basically rdfs:seeAlso, except he defines it as 'better info available from ...' rather than 'maybe some further info available ...'
DanC_jam: Megginson beats the "worse is better" drum pretty hard
evlist: Patrick Garvey explained the motivation of model based development for XML applications and mentions previous work on generating forms from schemas.
evlist: Their work is based on XSLT transformations generating XForms forms out of annotated W3C XML Schema schemas.
evlist: Patrick Garvey goes through the details of the pain of processing WXS schemas with XSLT.
evlist: Now comes demo time (William J. French)
evlist: More on their web page.
evlist: Their work is based on XSLT transformations generating XForms forms out of annotated W3C XML Schema schemas.
evlist: Patrick Garvey goes through the details of the pain of processing WXS schemas with XSLT.
evlist: Now comes demo time (William J. French)
evlist: More on their web page.
kendallclark: I pop up AIM rendezvous chat, to see who's around, and it's just me and danbri, who happens to be sitting next to me, so we hardly need it. Oh well.
evlist: The subject seems very popular and the rooms fills up.
kendallclark: A presentation motivated by the speaker's desire to show someone up, as he put it. :>
kendallclark: 3-D SVG gets around the 'visualization ghetto' problem of VRML/X3D
kendallclark: Uses something called the "world3D" XML vocabulary
kendallclark: projection/lighting computed by client javascript
kendallclark: A presentation motivated by the speaker's desire to show someone up, as he put it. :>
kendallclark: 3-D SVG gets around the 'visualization ghetto' problem of VRML/X3D
kendallclark: Uses something called the "world3D" XML vocabulary
kendallclark: projection/lighting computed by client javascript
edd: David Webber and Paul Stannard, Smartdraw. Another product presentation.
edd: Creating document pipelines and process-oriented applications through diagramming.
edd: Creating document pipelines and process-oriented applications through diagramming.
evlist: SVG export filters are not enough, native SVG editors that focus on design objective are needed.
evlist: Styling graphics through CSS is something really novel. Lets you style many graphics with a single CSS.
evlist: Or style a single SVG with multiple CSS for multiple medias.
evlist: Demo styling the same SVG for a poster, a small image and a fax with 3 different CSS.
evlist: Demo of diagrams generated through XSLT. Pretty impressive, I'd be curious to see the XSLT!
evlist: Styling graphics through CSS is something really novel. Lets you style many graphics with a single CSS.
evlist: Or style a single SVG with multiple CSS for multiple medias.
evlist: Demo styling the same SVG for a poster, a small image and a fax with 3 different CSS.
evlist: Demo of diagrams generated through XSLT. Pretty impressive, I'd be curious to see the XSLT!
evlist: Presenter(s): Barbara Hutson, Desk Officer, EPA, United States
evlist: Chris Lilley says they can't locate our speaker... If anyone has seen her, please wave!
evlist: This talk is cancelled...
evlist: Chris Lilley says they can't locate our speaker... If anyone has seen her, please wave!
evlist: This talk is cancelled...
edd: Jeff Barr, Amazon.com. Title updated since abstract submitted.
edd: Part of the product presentation track, so expect biased presentation.
edd: Explains Amazon customer sets: buyers, sellers (merchants who sell on Amazon's platform), Web site owners (Associates), and Developers (people who use their web services)
edd: Amazon Associates founded in 1996, over 1,000,000 registered associates.
edd: They've developed many feature components inside their firewall as part of their platform. They're just starting to expose some of these web services to customers.
edd: These components include things like the catalog, shopping cart, personalization
edd: Says they have ambitious plans to expose many more web services.
kendallclark: Amazon should provide a UPC barcode sticker with each book purchased, so I can track my library. (Thanks to Bijan for suggesting this to me the other day, whilst browsing my stacks.)
edd: How Amazon got to web services: partners needed better data access -- some had XML feeds, others scraped -- this was expensive and brittle method of collaboration.
edd: Web service aims: support industry stds, provide remote access to data and functionality, decouple presentation, create a software platform, unlock creatiity, and leverage further their internal investment in their web sites.
edd: Planning for web services, issues: Revenue -- intended to make their web services strategy revenue generating, rather than a "science experiment." So they'd support Associates and Sellers as their first stage. Licence Issues -- provide a degree of openness, protect rights of developers and Amazon, ensure data freshness, support Amazon's business model, and control their server load.
edd: Licensing issues resolved by constraints: 1 call per second, cannot resell data, must link to amazon.com, store non-pricing data for up to 24 hours, pricing data for up to 1 hour
kendallclark: (Does that mean pricing data changes rapidly, as a rule?)
edd: Pricing data includes availability data, which I think is the key issue.
edd: Protocol decisions: should they support SOAP or XML over HTTP (REST). They decided to let developers make the choice and support both. Though SOAP is the industry standard, it makes up just 15% of calls to the Amazon web services.
kendallclark: Awesome statistic. :>
edd: Noted the possibility of XSLT transformations as being a factor in the more vanilla XML delivery: they'll even do the XSLT transform for you.
edd: Issue: how to create a platform for developers. Resolution: use best practices from normal software world: document APIs, commit to API stability and backward compatibility.
edd: Developer support: how to help developers to succeed. Resolution: discussion board, weekly developer chats, regular newsletter, frequent releases, online FAQ.
edd: (Maybe some hints here for the semweb world in engaging developers.)
edd: "Responsivity" and openness to their developer community has worked well for them.
edd: History of releases: July 2003 AWS 1.0 (SOAP/REST interfaces, SDK, merchandising), Nov 2002 v 2.0 (market place support, quick click), April 2003, v3.0 (seller APIs, remote shopping cart, UK data), July 2003 incorp German and Japanese data.
edd: Growing their internal developer team by 5 times for next year -- pretty good marker of how they're getting behind the strategy.
edd: Moving out of purely web sites: there's an add-on for Office 2003 that can search the catalog and insert product metadata into your documents in various forms.
kendallclark: So, how long till MS feels "threatened" by Amazon and buys it? :>
edd: Challenge to developers using Amazon Web Services: got to know XML, XSLT, understand architecture of a distributed system, deploy caching for decent performance. Log the results of the calls, handle error conditions, degrade gracefully.
edd: (Room pretty packed out here.)
edd: Techs underneath: Linux, Apache, Perl, C++, Oracle
edd: Part of the product presentation track, so expect biased presentation.
edd: Explains Amazon customer sets: buyers, sellers (merchants who sell on Amazon's platform), Web site owners (Associates), and Developers (people who use their web services)
edd: Amazon Associates founded in 1996, over 1,000,000 registered associates.
edd: They've developed many feature components inside their firewall as part of their platform. They're just starting to expose some of these web services to customers.
edd: These components include things like the catalog, shopping cart, personalization
edd: Says they have ambitious plans to expose many more web services.
kendallclark: Amazon should provide a UPC barcode sticker with each book purchased, so I can track my library. (Thanks to Bijan for suggesting this to me the other day, whilst browsing my stacks.)
edd: How Amazon got to web services: partners needed better data access -- some had XML feeds, others scraped -- this was expensive and brittle method of collaboration.
edd: Web service aims: support industry stds, provide remote access to data and functionality, decouple presentation, create a software platform, unlock creatiity, and leverage further their internal investment in their web sites.
edd: Planning for web services, issues: Revenue -- intended to make their web services strategy revenue generating, rather than a "science experiment." So they'd support Associates and Sellers as their first stage. Licence Issues -- provide a degree of openness, protect rights of developers and Amazon, ensure data freshness, support Amazon's business model, and control their server load.
edd: Licensing issues resolved by constraints: 1 call per second, cannot resell data, must link to amazon.com, store non-pricing data for up to 24 hours, pricing data for up to 1 hour
kendallclark: (Does that mean pricing data changes rapidly, as a rule?)
edd: Pricing data includes availability data, which I think is the key issue.
edd: Protocol decisions: should they support SOAP or XML over HTTP (REST). They decided to let developers make the choice and support both. Though SOAP is the industry standard, it makes up just 15% of calls to the Amazon web services.
kendallclark: Awesome statistic. :>
edd: Noted the possibility of XSLT transformations as being a factor in the more vanilla XML delivery: they'll even do the XSLT transform for you.
edd: Issue: how to create a platform for developers. Resolution: use best practices from normal software world: document APIs, commit to API stability and backward compatibility.
edd: Developer support: how to help developers to succeed. Resolution: discussion board, weekly developer chats, regular newsletter, frequent releases, online FAQ.
edd: (Maybe some hints here for the semweb world in engaging developers.)
edd: "Responsivity" and openness to their developer community has worked well for them.
edd: History of releases: July 2003 AWS 1.0 (SOAP/REST interfaces, SDK, merchandising), Nov 2002 v 2.0 (market place support, quick click), April 2003, v3.0 (seller APIs, remote shopping cart, UK data), July 2003 incorp German and Japanese data.
edd: Growing their internal developer team by 5 times for next year -- pretty good marker of how they're getting behind the strategy.
edd: Moving out of purely web sites: there's an add-on for Office 2003 that can search the catalog and insert product metadata into your documents in various forms.
kendallclark: So, how long till MS feels "threatened" by Amazon and buys it? :>
edd: Challenge to developers using Amazon Web Services: got to know XML, XSLT, understand architecture of a distributed system, deploy caching for decent performance. Log the results of the calls, handle error conditions, degrade gracefully.
edd: (Room pretty packed out here.)
edd: Techs underneath: Linux, Apache, Perl, C++, Oracle
kendallclark: David Burdett, presenter
kendallclark: A big review of WS standards
kendallclark: Why Standards? interop & lower cost
kendallclark: the typical 19th c. railway gauge example (wow, that's so tired)
kendallclark: "standardization of 'net protocols => explosive growth of the web"
kendallclark: standardization of WS will lead to "explosive biz integration growth" in the next 20 yrs.
kendallclark: Yet another WS stack -- hasn't all this been rather covered before?
kendallclark: It's very late '02.
kendallclark: Complex WS 'provide for secure reliable delivery of messages'
kendallclark: as opposed to simple WS, one wonders a bit snarkily.
kendallclark: ws-reliability and ws-reliable messaging -- competing specs (?)
kendallclark: Hmm, he does make an interesting point in the new way that standards are being developed; small grp of vendors works something out privately, then tries to have a standards body bless it. (This is really extra-W3C efforts, I think. That's harder to do in W3C.)
kendallclark: The data-related WS standards are particularly mature.
kendallclark: WSDL seems to have won the day for describing interfaces; BEPL for "internal process defs" and ws-choreography for multi-org cooperation.
kendallclark: after data and messaging/coordination layers, he turns to 'semantic definitions' -- i.e., how do we specify 'order', 'invoice', etc.
kendallclark: Says UBL (universal biz lang) 'gaining traction'
kendallclark: Next, 'profiles, policies, agreements' (this is the level at which a SemWeb company like network inference wants to get involved)
kendallclark: Next, need a way to manage these WS.
kendallclark: A big review of WS standards
kendallclark: Why Standards? interop & lower cost
kendallclark: the typical 19th c. railway gauge example (wow, that's so tired)
kendallclark: "standardization of 'net protocols => explosive growth of the web"
kendallclark: standardization of WS will lead to "explosive biz integration growth" in the next 20 yrs.
kendallclark: Yet another WS stack -- hasn't all this been rather covered before?
kendallclark: It's very late '02.
kendallclark: Complex WS 'provide for secure reliable delivery of messages'
kendallclark: as opposed to simple WS, one wonders a bit snarkily.
kendallclark: ws-reliability and ws-reliable messaging -- competing specs (?)
kendallclark: Hmm, he does make an interesting point in the new way that standards are being developed; small grp of vendors works something out privately, then tries to have a standards body bless it. (This is really extra-W3C efforts, I think. That's harder to do in W3C.)
kendallclark: The data-related WS standards are particularly mature.
kendallclark: WSDL seems to have won the day for describing interfaces; BEPL for "internal process defs" and ws-choreography for multi-org cooperation.
kendallclark: after data and messaging/coordination layers, he turns to 'semantic definitions' -- i.e., how do we specify 'order', 'invoice', etc.
kendallclark: Says UBL (universal biz lang) 'gaining traction'
kendallclark: Next, 'profiles, policies, agreements' (this is the level at which a SemWeb company like network inference wants to get involved)
kendallclark: Next, need a way to manage these WS.
MarkB: my day job for the past few months
edd: Second XML 2003 keynote. Shantanu Narayen, Adobe.
edd: Hmm, the scheduled title is "Extend the Power of Business Processes: Bringing Data and Presentation Together with XML"
kendallclark: "intelligent documents": store xml, store routing info (so that it can "move"), store presentation ("accessible for the humans", as opposed to the robots, of course)
edd: Intelligent Documents contain XML Documents, XML Forms and Metadata, according to Adobe.
edd: Notably under "XML Forms" doesn't cite XForms, but JavaScript and XSLT instead.
edd: Under Metadata they say XMP and RDF.
edd: The XML Documents bit is "templates", XSLT, XSL-FO, SVG
kendallclark: He hasn't said anything, near as I can tell, about what representation the document routing info is in...? XML and Schema, I guess.
edd: So seems like a business document in their view bundles presentation, behaviour, business logic (routing, etc.) as well as content. PDF is the vehicle for this, but with XML smarts inside.
edd: Adobe geting involved with ebXML, acquired Yellow Dragon.
edd: So I wonder if they see PDFs being passed around with ebXML...?
kendallclark: Yes, seems so, in a rather big way.
edd: Using ebXML registry as a repository for intelligent business forms
edd: The XML documents can be filled out through Adobe Acrobat Reader.
edd: The documents look like the pieces of paper they replace, but have inbuilt validation and business logic.
edd: Digital signatures once the form is filled out.
edd: My big doubt about this is that paper emulation seems a bit limiting.
edd: The resulting XML document can be generated to a specified schema. XDP, XML data package, seems to be the tech they use to wrap it all up together (glad to see someone did something about packaging :)
kendallclark: Interestingly, when I wrote about XML tools in regulatory/gov't environments back in Sept, Adobe had a very tiny presence. One wonders whether that's changing.
edd: Showing "Adobe Designer" the editing tool for these forms. JavaScripting can be used for validation, etc. The UI is described in an XML language, I couldn't see the namespace quickly enough, but he shows the XML source behind the PDF prettiness.
edd: The data in the document can be bound in various ways: to an XML Schema, to a relational database, as inputs to web services. Done via drag and drop linking.
edd: Starting from blank, demo shows that a schema element can be dropped onto a blank form and a first-pass UI is automatically generated, which can be prettified.
edd: The fact that these documents can be bound to static documents, web services, databases is good: seems to provide for a multiplicity of routing options.
kendallclark: Adobe is excited about this "new technology" and it hopes that I'm excited to! I'll have to think about that.
edd: I have reservations about the paper-emulation aspect of this, but maybe that's not such a problem.
edd: Now talking about working for interoperability with SAP, IBM and others etc to allow their frontends to work with the major business backends.
edd: Hmm, the scheduled title is "Extend the Power of Business Processes: Bringing Data and Presentation Together with XML"
kendallclark: "intelligent documents": store xml, store routing info (so that it can "move"), store presentation ("accessible for the humans", as opposed to the robots, of course)
edd: Intelligent Documents contain XML Documents, XML Forms and Metadata, according to Adobe.
edd: Notably under "XML Forms" doesn't cite XForms, but JavaScript and XSLT instead.
edd: Under Metadata they say XMP and RDF.
edd: The XML Documents bit is "templates", XSLT, XSL-FO, SVG
kendallclark: He hasn't said anything, near as I can tell, about what representation the document routing info is in...? XML and Schema, I guess.
edd: So seems like a business document in their view bundles presentation, behaviour, business logic (routing, etc.) as well as content. PDF is the vehicle for this, but with XML smarts inside.
edd: Adobe geting involved with ebXML, acquired Yellow Dragon.
edd: So I wonder if they see PDFs being passed around with ebXML...?
kendallclark: Yes, seems so, in a rather big way.
edd: Using ebXML registry as a repository for intelligent business forms
edd: The XML documents can be filled out through Adobe Acrobat Reader.
edd: The documents look like the pieces of paper they replace, but have inbuilt validation and business logic.
edd: Digital signatures once the form is filled out.
edd: My big doubt about this is that paper emulation seems a bit limiting.
edd: The resulting XML document can be generated to a specified schema. XDP, XML data package, seems to be the tech they use to wrap it all up together (glad to see someone did something about packaging :)
kendallclark: Interestingly, when I wrote about XML tools in regulatory/gov't environments back in Sept, Adobe had a very tiny presence. One wonders whether that's changing.
edd: Showing "Adobe Designer" the editing tool for these forms. JavaScripting can be used for validation, etc. The UI is described in an XML language, I couldn't see the namespace quickly enough, but he shows the XML source behind the PDF prettiness.
edd: The data in the document can be bound in various ways: to an XML Schema, to a relational database, as inputs to web services. Done via drag and drop linking.
edd: Starting from blank, demo shows that a schema element can be dropped onto a blank form and a first-pass UI is automatically generated, which can be prettified.
edd: The fact that these documents can be bound to static documents, web services, databases is good: seems to provide for a multiplicity of routing options.
kendallclark: Adobe is excited about this "new technology" and it hopes that I'm excited to! I'll have to think about that.
edd: I have reservations about the paper-emulation aspect of this, but maybe that's not such a problem.
edd: Now talking about working for interoperability with SAP, IBM and others etc to allow their frontends to work with the major business backends.
edd: Udell talking about blogs, XPath, RSS, context.
edd: Mentions Microsoft's InfoPath for XML document flow.
edd: MS dropped the ball in not allowing XML document creation from Outlook, precisely the place they should have put it.
edd: His general thrust is that the communications methods by which we conduct most business, email and IM, are still unstructured.
edd: CSS could be a backdoor for getting metadata into XHTML: people care much more about appearance often than putting metadata in. Promote cool-looking styles as a backdoor for metadata.
edd: Mentions Microsoft's InfoPath for XML document flow.
edd: MS dropped the ball in not allowing XML document creation from Outlook, precisely the place they should have put it.
edd: His general thrust is that the communications methods by which we conduct most business, email and IM, are still unstructured.
edd: CSS could be a backdoor for getting metadata into XHTML: people care much more about appearance often than putting metadata in. Promote cool-looking styles as a backdoor for metadata.
libby: a simple experimental perl scrpt that generates rdfical from an rss file
libby: it expects a title that's something like "title, date, location", comma seprated. It tries to look up the location too.
libby: it also saves the file
libby: not sure whether to add creation and saving to a bot, sorta redoing the chump, or just set up a crontab or something
libby: my bot knows about events like this via a crontab - try '!bydate 2004'
libby: some rough ideas about using weblogs and the chump (that sort of work)
libby: it expects a title that's something like "title, date, location", comma seprated. It tries to look up the location too.
libby: it also saves the file
libby: not sure whether to add creation and saving to a bot, sorta redoing the chump, or just set up a crontab or something
libby: my bot knows about events like this via a crontab - try '!bydate 2004'
libby: some rough ideas about using weblogs and the chump (that sort of work)
Jhendler: some of my students playing with generation of forms from OWL -- looks to be easier than XML forms according to them
Jhendler: not clear, however, if as general